MYSM1-AR Complex-Mediated Repression of Akt/C-Raf/GSK-3Β Signaling Impedes Castration-Resistant Prostate Cancer Growth

www.aging-us.com AGING 2019, Vol. 11, No. 22

Research Paper MYSM1-AR complex-mediated repression of Akt/c-Raf/GSK-3β signaling impedes castration-resistant prostate cancer growth

Jinbo Sun1,*, Xiangnan Hu1,*, Yongheng Gao2,*, Qisheng Tang1, Zhining Zhao1,3, Wenjin Xi4, Fan Yang1, Wei Zhang1, Yue Song5, Bin Song1, Tao Wang4,6, He Wang1

1Department of Urology, Tangdu Hospital, Fourth Military Medical University, Xi'an, Shaanxi 710038, China 2Department of Respiratory and Critical Care Medicine, Tangdu Hospital, Fourth Military Medical University, Xi'an, Shaanxi 710038, China 3Clinical Laboratory, 451 Hospital of Chinese People's Liberation Army, Xi'an, Shaanxi 710054, China 4State Key Laboratory of Cancer Biology, Department of Immunology, Fourth Military Medical University, Xi’an, Shaanxi 710032, China 5Department of Orthopedics, Xijing Hospital, Fourth Military Medical University, Xi’an, Shaanxi 710032, China 6Department of Medical Genetics and Developmental Biology, Fourth Military Medical University, Xi’an, Shaanxi 710032, China *Equal contribution

Correspondence to: He Wang, Tao Wang, Bin Song; email: [email protected], [email protected], [email protected] Keywords: castration-resistant prostate cancer, growth, MYSM1, AR, Akt Received: August 20, 2019 Accepted: November 8, 2019 Published: November 24, 2019

Copyright: Sun et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

ABSTRACT

Epigenetic alterations that lead to dysregulated gene expression in the progression of castration-resistant prostate cancer (CRPC) remain elusive. Here, we investigated the role of histone deubiquitinase MYSM1 in the pathogenesis of prostate cancer (PCa). Tissues and public datasets of PCa were evaluated for MYSM1 levels. We explored the effects of MYSM1 on cell proliferation, senescence and viability both in vitro and in vivo. Integrative database analyses and co-immunoprecipitation assays were performed to elucidate genomic association of MYSM1 and MYSM1-involved biological interaction network in PCa. We observed that MYSM1 were downregulated in CRPC compared to localized prostate tumors. Knockdown of MYSM1 promoted cell proliferation and suppressed senescence of CRPC cells under condition of androgen ablation. MYSM1 downregulation enhanced the tumorigenic ability in nude mice. Integrative bioinformatic analyses of the significantly associated genes with MYSM1 revealed MYSM1-correlated pathways, providing substantial clues as to the role of MYSM1 in PCa. MYSM1 was able to bind to androgen receptor instead of increasing its expression and knockdown of MYSM1 resulted in activation of Akt/c-Raf/GSK-3β signaling. Together, our findings indicate that MYSM1 is pivotal in CRPC pathogenesis and may be established as a potential target for future treatment.

INTRODUCTION androgen synthesis or androgen receptor (AR) activity [2, 3], remain the mainstay treatment for locally advanced Prostate cancer (PCa) is one leading cause of cancer- or metastatic PCa. Unfortunately, the majority of patients related death among men in Western countries [1]. who have undergone ADT, even initially effective in Although surgical intervention has been shown to be tumor regression, eventually become resistant to the efficacious in eradicating localized PCa, androgen treatment [4] and progress to be castration-resistant PCa deprivation therapy (ADT), aimed at suppressing (CRPC). Castration resistance represents an enormous www.aging-us.com 10644 AGING clinical challenge and improvements in therapies as well In this study, we investigated the role of MYSM1 in as causative biomarkers for CRPC are in urgent need. carcinogenesis and progression of PCa. We revealed the decreased expression of MYSM1 and confirmed its AR with transcription activity is expressed in CRPC and tumor-suppressing functions in CRPC. Mechanistically, may play an important role in the setting of castration MYSM1 may exert its effect through interplay with AR via intratumoral androgen biosynthesis or interplay with and inhibiting the activation of PI3K/Akt signaling. This other growth-promoting and prosurvival pathways [5, 6]. work suggests that MYSM1 can be functionally essential The heterogeneity in AR expression or activity suggest for CRPC progression. that AR deficiency may be proposed as a potential way in which prostate cancer cells escape androgen RESULTS deprivation therapy [7], with compensatory signaling pathways activated concomitantly. The PI3K/Akt MYSM1 is downregulated in castration-resistant pathway is a well-established oncogenic pathway in prostate cancer and inversely correlated with human cancer [8–10] and is involved in resistance to progression of prostate cancer AR-targeted therapy in prostate cancer [11–13]. Alterations of nodes in PI3K/Akt pathway are reported To explore the expression of MYSM1 in human tumors, to occur in 42% of primary PCa and 70% of metastatic we performed web-based data mining to analyze tumors [14]. PTEN loss is frequently observed during The Cancer Genome Atlas (TCGA) datasets via PCa progression, especially in advanced prostate tumors GEPIA bioinformatics. We found that MYSM1 mRNA [15–17]. Thus, the PI3K/Akt pathway, which is levels were downregulated in ACC (adrenocortical antagonized by tumor suppressor PTEN, is constitutively carcinoma), CESC (cervical squamous cell carcinoma activated in prostate cancers with PTEN deficiency, and endocervical adenocarcinoma), COAD (colon leading to enhanced tumor cell survival, metastasis and adenocarcinoma), LUAD (lung adenocarcinoma), LUSC castration-resistant growth [12, 18, 19]. In addition, (lung squamous cell carcinoma), OV (ovarian serous reciprocal feedback regulation between PI3K/Akt and cystadenocarcinoma), READ (rectum adenocarcinoma), AR signaling has been identified as a potent mechanism SKCM (skin cutaneous melanoma), THCA (thyroid of CRPC progression and a crucial issue for carcinoma), UCEC (uterine corpus endometrial monotherapies targeting AR or PI3K/Akt pathways [20– carcinoma) and UCS (uterine carcinosarcoma) compared 22]. Combination therapy cotargeting PI3K/Akt and AR with normal tissues (Figure 1A, Supplementary Figure signaling leads to significant regression of prostate 1). In addition, MYSM1 expression was higher only in cancer when compared with monotherapies [23–25], DLBC (lymphoid neoplasm diffuse large B-cell suggesting a coordinative role in supporting tumor lymphoma) and THYM (thymoma) than that in normal survival. tissues (Figure 1A, Supplementary Figure 2). The expression of MYSM1 was decreased in PRAD (prostate Myb-like SWIRM and MPN domains 1 (MYSM1) acts adenocarcinoma) compared with that in normal prostate as a histone H2A deubiquitinase and is responsible for glands, but the change was not statistically significant removal of ubiquitin from monoubiquitinated histone (Figure 1A, 1B). Immunostaining analyses revealed that H2A at lysine 119 (H2AK119ub) [26]. Early studies MYSM1 protein levels were a little lower in prostate have linked MYSM1 to hematopoiesis where MYSM1 cancers when compared with benign prostatic modulates hematopoietic stem cell function and survival hyperplasia (Figure 1C). [27–29]. By coordinating histone modifications and transcription factors recruitment, MYSM1 plays a To further verify the expression of MYSM1 in prostate critical role in control of lymphocyte differentiation and cancer, we analyzed 2 microarray datasets (Grasso tissue development [29, 30]. Additionally, it has been Prostate and Taylor Prostate 3) from Oncomine reported that MYSM1 functions as a central negative database. Similarly, no significant differences of regulator of inflammatory response and immune system MYSM1 levels were observed between prostate cancers to prevent excessive inflammation and self-destructive and prostate glands (Figure 1D). However, we found immune response [31, 32]. Little is known about the role that compared with localized PCa patients, the of MYSM1 in human cancers. Few studies found that expression of MYSM1 was significantly downregulated MYSM1 was involved in melanoma growth and in metastatic castration-resistant prostate cancer colorectal cancer metastasis [33, 34]. It is worth noting (CRPC) patients (Figure 1E). Moreover, Oncomine that MYSM1 was reported to participate in regulation of analyses also revealed the clinically relevant signatures AR-dependent gene transcription in PTEN-deficient of MYSM1 expression in human PCa progression LNCaP cells [26]. However, the prostate-specific role of (Table 1). Although MYSM1 expression did not MYSM1, particularly its role in CRPC, has not been significantly correlate with age, PSA level, T stage, explored in detail. extracapsular extension, seminal vesicle involvement, www.aging-us.com 10645 AGING surgical margins status, hormone therapy, recurrence-free survival in PCa patients (Figure 1J, 1K). chemotherapy and radiotherapy (Table 1, P > 0.05), the The reason for insignificant correlation of MYSM1 level of MYSM1 was inversely associated with Gleason mRNA level with prognosis may be due in part to grade, Gleason score, N stage and recurrence status insufficient sample size. Taken together, these results (Figure 1F–1I). Kaplan-Meier analyses of Taylor suggest that dysregulation of MYSM1 may play a Prostate 3 cohort indicated no significant associations significant role in PCa progression and contribute to of MYSM1 expression with overall survival and development of castration resistance.

Figure 1. MYSM1 is downregulated in castration-resistant prostate cancer and inversely correlated with progression of prostate cancer. (A) Pan-cancer analyses for mean expression levels of MYSM1 in different types of cancers. (B) MYSM1 mRNA levels in prostate cancers (T = tumor, N = normal, PRAD = prostate adenocarcinoma). Data (A–B) were acquired from TCGA database and analyzed via GEPIA bioinformatics. (C) Representative IHC staining images of MYSM1 in benign prostatic hyperplasia (BPH) and prostate cancer (PCa) tissues. Scale bars are 20 μm. (D–E) MYSM1 expression levels based on Grasso Prostate dataset from Oncomine database. (F–I) MYSM1 is differentially expressed in prostate cancer patients and data acquired from Taylor Prostate 3 cohort are analyzed based on Gleason grade (F), Gleason score (G), N stage (H) and recurrence status (I) via Oncomine database. Data are shown as mean ± SD. n.s = no significance, * P < 0.05 and ** P < 0.01 (Student’s t-test). (J–K) Correlation of MYSM1 with overall survival (J) and recurrence-free survival (K) in prostate cancer patients (n = 140). Data were acquired from Taylor Prostate 3 cohort via Oncomine database. Log-rank test was applied to determine the significance levels. www.aging-us.com 10646 AGING Table 1. Demographic and Clinicopathological Characteristics of Prostate Cancer Patients (Taylor Prostate 3 Cohort) and Association of MYSM1 Expression with Clinicopathological Parameters (Chi-square test). MYSM1 mRNA expression Clinicopathological parameters Frequency (%) P value < median ≥median Age (n=150) < 60 93 (62.0) 43 50 0.239 ≥ 60 57 (38.0) 32 25 Pre-diagnosis biopsy PSA (n=147) < 10 ng/ml 115 (78.2) 59 56 0.450 ≥ 10 ng/ml 32 (21.8) 14 18 Pre-treatment PSA (n=147) < 10 ng/ml 105 (71.4) 55 50 0.297 ≥ 10 ng/ml 42 (28.6) 18 24 T stage (n=141) T1-T2 86 (61.0) 40 46 0.352 T3-T4 55 (39.0) 30 25 Extracapsular extension (n=141) None 43 (30.5) 19 24 0.361a Capsular invasion 47 (33.3) 23 24 Focal 7 (5.0) 2 5 Established 44 (31.2) 26 18 Seminal vesicle involvement (n=141) Negative 119 (84.4) 58 61 0.617 Positive 22 (15.6) 12 10 Surgical margins (n=141) Negative 108 (76.6) 51 57 0.298 Positive 33 (23.4) 19 14 Hormone therapy (n=150) No 115 (76.7) 55 60 0.334 Yes 35 (23.3) 20 15 Chemotherapy (n=150) No 136 (90.7) 65 71 0.092 Yes 14 (9.3) 10 4 Radiotherapy (n=150) No 126 (84.0) 62 64 0.656 Yes 24 (16.0) 13 11 aFisher exact test.

MYSM1 knockdown promotes proliferation and control shRNA (shNC). The efficient knockdown of suppresses senescence of CRPC cells MYSM1 was confirmed by measuring MYSM1 expression at the mRNA (Figure 2A) and protein (Figure The observation that the MYSM1 expression decreases 2B) levels. Cells were cultivated in RPMI-1640 medium as the tumor develops castration resistance led us to supplemented with charcoal-stripped serum to mimic the evaluate the role of MYSM1 in CRPC. To determine the hormone-starvation conditions. MTT and flow putative function of MYSM1 in castration-resistant cytometry assays were performed to investigate the growth of PCa, we downregulated MYSM1 levels in influence of MYSM1 knockdown on proliferation and androgen-independent C4-2 and 22Rv1 cell lines using cell cycle in C4-2 and 22Rv1 cells stably expressing lentivirus MYSM1 shRNAs (shMYSM1) and negative shRNA targeting MYSM1 or negative control. We

www.aging-us.com 10647 AGING found that MYSM1 silencing in CRPC cells 3C). To further confirm the functional impact of significantly increased the proliferation as shown in MYSM1 on tumor growth, we performed IHC staining MTT assays (Figure 2C). Similarly, a significant change to analyze tumor tissues from nude mice. Our results of cell cycle distribution was detected in MYSM1- showed that the protein level of Ki-67, a marker for deficient cells. There was a decrease in the percentage of proliferation, was significantly upregulated in tumor G1-phase cells and an increase in that of S-phase cells sections from the 22Rv1/shMYSM1 group compared (Figure 2D), indicating an accelerated progression of cell with those from the 22Rv1/shNC group (Figure 3D). cycle. Because it has been reported that promoted cell Consistently, IHC staining for p-Akt, constitutively growth and cell cycle progression might be mediated by activated in CRPC cells, showed that downregulation of suppression of senescence and apoptosis, we next MYSM1 resulted in a significant increase in p-Akt assessed whether MYSM1 deletion in CRPC cells would expression in tumor tissues derived from mice implanted inhibit senescence and apoptosis induction. Therefore, with 22Rv1/shMYSM1 cells (Figure 3D). Together, we performed SA-β-gal staining assays to evaluate these data indicate that MYSM1 negatively regulates cellular senescence phenotype. Our results showed that CRPC growth in vivo. MYSM1 downregulation led to decreased proportion of β-gal positive cells (Figure 2E). Moreover, cells Integrative multi-omics analysis of MYSM1 mRNA transfected with siRNAs against MYSM1 and NC for expression in prostate cancers 48h were subjected to apoptosis analysis via Annexin V/PI-labeling flow cytometry. However, our results The LinkFinder module of LinkedOmics database was showed that MYSM1 depletion did not exert used to explore the association between MYSM1 mRNA considerable influence on apoptosis in CRPC cells and other gene expression by analyzing the RNA (Supplementary Figure 3). Resistance to senescence or sequencing data from TCGA PCa patient cohort (n = apoptosis has been proposed as a strategy for cancer cell 497). As indicated by the volcano plot (Figure 4A), 6624 survival and tumor growth promotion. In agreement with genes (dark red dots) displayed significant positive our results, prior research has shown that some cells are correlation with MYSM1 expression, whereas 6088 more susceptible to senescence rather than apoptosis genes (dark green dots) showed strong negative even after undergoing extensive extrinsic stimuli [35]. association (FDR < 0.01). The top 50 most significant Further analyses of PCa patients from LinkedOmics genes positively and negatively associated with MYSM1 database revealed a negative correlation between mRNA level were visualized in the heat maps (Figure MYSM1 mRNA and gene transcripts related to 4B). These findings suggest a widespread influence of proliferation and cell cycle, including CDK4, CCND3 MYSM1 on the transcriptome. The LinkFinder also and CCNE1 (Figure 2F). Moreover, we observed that output statistical scatter plots for individual genes. MYSM1 transcription was strongly associated with the MYSM1 gene transcript level demonstrated a strong expression of tumor suppressor RB1 in PCa patients positive association with the expression of LCOR (Figure 2F). Collectively, these data indicate that (positive rank #1, r = 0.84, P = 3.94e-136) and SMG1 MYSM1 expression in CRPC cells results in the (positive rank #2, r = 0.83, P = 4.30e-130) (Figure 4C). suppression of androgen-independent growth and Both LCOR and SMG1 protein are known to function as induction of cell cycle arrest as well as cellular a tumor suppressor in multiple cancers and mediate the senescence, suggesting a tumor-suppressive role of repression of tumor growth by regulating oncogenic MYSM1 in CRPC cells. proteins involved in cell cycle [36, 37]. Using Cancer Regulome tools, further analyses of TCGA PCa patients Downregulation of MYSM1 facilitates prostate were performed to explore the genome-level associations cancer growth in vivo of MYSM1 expression with various molecular features, including gene expression, DNA methylation, somatic To further determine the anti-tumor role of MYSM1 in copy number, microRNA expression, somatic mutation CRPC growth in vivo, we subcutaneously engrafted and protein level (RPPA). The significant associations nude mice with 22Rv1 cells stably expressing shRNA between MYSM1 level and gene expression features targeting either MYSM1 or negative control and tracked within the context of genomic coordinates were presented tumor growth. Our results showed that the tumors as a circular graph (Figure 4D) and a network (Figure formed in mice from 22Rv1/shMYSM1 group were 4E). Consistent with the results in Figure 4A–4C, the significantly larger than those from 22Rv1/shNC group transcription levels of LCOR and SMG1 (red) were (Figure 3A). Moreover, we observed that the inhibition significantly correlated with MYSM1 mRNA expression of MYSM1 significantly affected tumor growth in nude (Figure 4E). We next performed gene ontology (GO) and mice (Figure 3B). In addition, tumor weights of 22Rv1 KEGG pathway analyses of genes significantly xenografts in immunodeficient mice were significantly associated with MYSM1 expression using WebGestalt increased by downregulating MYSM1 levels (Figure via LinkedOmics bioinformatics (Figure 4F). Our results

www.aging-us.com 10648 AGING indicated that the proteins encoded by these genes are activity (FDR < 0.0001). KEGG pathway analysis localized mainly in cytoskeleton, cell projection and revealed that MYSM1 mRNA was highly associated with endoplasmic reticulum (FDR < 0.0001). They are genes involved in metabolic pathways and mTOR significantly enriched in cellular response to stress, cell signaling pathway (FDR < 0.01). Taken together, these cycle and negative regulation of gene expression (FDR < findings suggest a potential role of MYSM1 in the 0.0001) and participate primarily in nucleotide binding, biological interaction network correlated with tumor transcription regulator activity and enzyme regulator progression.

Figure 2. MYSM1 knockdown promotes proliferation and suppresses senescence of prostate cancer cells. (A–B) MYSM1 expression levels in CRPC cell lines (C4-2, 22Rv1) stably expressing shRNA targeting MYSM1 (shMYSM1) or negative control (shNC) were detected by qRT-PCR (A) and Western blot (B) analyses. Data are shown as mean ± SEM of 3 replicates. ** P < 0.01 and *** P < 0.001 (one- way ANOVA test). (C) Proliferation of shNC/shMYSM1-treated C4-2/22Rv1 cells was evaluated by MTT assay. Data are shown as mean ± SEM of 3 replicates. (D) Flow cytometry analysis of cell cycle in C4-2/22Rv1 cells treated with shNC/shMYSM1. Data are shown as mean ± SD of 3 replicates. ** P < 0.01 and *** P < 0.001 (Student’s t-test). (E) Representative images of SA-β-gal staining in C4-2/22Rv1 cells treated with shNC/shMYSM1. Scale bars are 25 μm. Data are shown as mean ± SD of 3 replicates. *** P < 0.001 and **** P < 0.0001 (Student’s t-test). (F) Correlation of MYSM1 with CDK4, CCND3, CCNE1 and RB1 mRNA levels in prostate cancers. Data were collected from TCGA database and analyzed via LinkedOmics bioinformatics. Pearson correlation coefficients and significance levels are indicated. www.aging-us.com 10649 AGING MYSM1 interacts with AR and inhibits activation of analyze AR expression in xenograft tumor tissues from Akt/c-Raf/GSK-3β signaling in prostate cancer nude mice (Figure 5C). However, we observed that MYSM1 knockdown led to no apparent changes in AR Considering the significant role of MYSM1 in CRPC, expression (Figure 5A–5C). Given the repression of AR we next sought to elucidate the molecular mechanisms transactivation by interaction of LCOR with AR [36], we by which MYSM1 suppressed CRPC progression. then concentrated on the interaction between MYSM1 Previous studies have shown that LCOR can act as a and AR to dissect the functional significance of MYSM1 corepressor for AR signaling [36] which is still in AR signaling. To verify the potential interaction indispensable for CRPC. As our results demonstrated a between MYSM1 and AR, we carried out protein co- strong association between MYSM1 and LCOR, we thus immunoprecipitation (Co-IP) assays with validated focused on investigating whether MYSM1 regulates AR antibodies against MYSM1 and AR in C4-2 cells with or signaling. To study the effect of MYSM1 on AR without siMYSM1 treatment. The Co-IP results expression, qRT-PCR (Figure 5A) and western blot confirmed the binding between MYSM1 and AR in (Figure 5B) analyses were performed to detect the CRPC cells (Figure 5D), which may be conducive to the expression levels of AR in MYSM1-downregulated activation of AR signaling. A reciprocal feedback has CRPC cells. In addition, we performed IHC staining to been reported to exist between AR and Akt signal

Figure 3. Downregulation of MYSM1 facilitates prostate cancer growth in vivo. (A) Representative image of tumors formed in nude mice bearing 22Rv1/shNC or 22Rv1/shMYSM1 cells. (B) The volumes of tumors derived from nude mice subcutaneously implanted with 22Rv1/shNC or 22Rv1/shMYSM1 cells were monitored for 30 days. Data are shown as mean ± SD of 3 replicates. (C) The animals were sacrificed 30 days after injection and tumor weights were evaluated. Data are shown as mean ± SD of 3 replicates. (D) Representative photographs of IHC staining of p-Akt and Ki-67 in tumor tissues from 22Rv1/shNC and 22Rv1/shMYSM1 nude mice groups. Scale bars are 30 μm. ** P < 0.01 and *** P < 0.001 (Student’s t-test).

www.aging-us.com 10650 AGING pathways [21, 22]. Bioinformatics analyses of a TCGA PHLPP1 are known to be key regulators in activation of PCa cohort via LinkedOmics indicted that MYSM1 Akt signaling. In addition, KEGG pathway analysis also significantly correlates with the expression levels of revealed a significant enrichment of MYSM1-associated PHLLP1 and PTEN (Figure 5E). Both PTEN and genes in mTOR signaling (Figure 4F). To further

Figure 4. Integrative multiomics analysis of MYSM1 mRNA expression in prostate cancers. (A) Association of MYSM1 with genes differentially expressed in prostate cancers. (B) Heatmaps for genes positively (left) and negatively (right) correlated with MYSM1 in prostate cancers (top 50). (C) Correlation of MYSM1 with LCOR (upper) and SMG1 (lower) mRNA levels in prostate cancers. Pearson correlation coefficients and significance levels are indicated. Data (A–C) were collected from TCGA database and analyzed via LinkedOmics bioinformatics. (D) Statistically significant associations of MYSM1 with genomic coordinates are indicated by arcs connecting pairs of dots in prostate cancers. (E) Network view of correlations between MYSM1 and other genes in prostate cancers. Data (D–E) were acquired from TCGA database and analyzed via Regulome Explorer. (F) Significantly enriched GO annotations and KEGG pathways of genes correlated with MYSM1 in prostate cancers. Data were acquired from TCGA database and analyzed via GSEA bioinformatics. www.aging-us.com 10651 AGING validate the link between MYSM1 and activation of Akt (Figure 5F). Western blot results further indicated that pathway, we performed qRT-PCR and western blot stable knockdown of MYSM1 led to increased levels of assays in C4-2 cells. The qRT-PCR analyses showed p-Akt(S473), p-c-Raf and p-GSK-3β without affecting that siRNA-mediated downregulation of MYSM1 the expression of p-Akt(T308), p-PDK1 and total resulted in a significant decrease in PHLPP1 expression, proteins including Akt, c-Raf and GSK-3β (Figure 5G). confirming the association of MYSM1 with PHLPP1 In summary, these findings demonstrate that MYSM1

Figure 5. MYSM1 interacts with AR and inhibits activation of Akt/c-Raf/GSK-3β signaling in prostate cancer. (A) qRT-PCR analyses for MYSM1 and AR mRNA levels following transfection of NC/siMYSM1 into C4-2 and 22Rv1 cells. Data are shown as mean ± SEM of 3 replicates. (B) Western blot analyses for MYSM1 and AR protein levels in C4-2 and 22Rv1 cells treated with shNC/shMYSM1. (C) Representative IHC staining images of AR in xenograft tumor tissues. Scale bars are 30 μm. (D) Co-IP assays of MYSM1 and AR in C4-2 cells with (down) or without (up) siMYSM1 treatment. (E) Correlation of MYSM1 with PHLPP1 (up) and PTEN (down) mRNA levels in prostate cancers. Pearson correlation coefficients and significance levels are indicated. Data were collected from TCGA database and analyzed via LinkedOmics bioinformatics. (F) qRT-PCR analyses for MYSM1 and PHLPP1 mRNA levels following transfection of NC/siMYSM1 into C4-2 cells. Data are shown as mean ± SEM of 3 replicates. (G) Western blot analyses of lysates from C4-2 cells treated with shNC/shMYSM1 were probed with the indicated antibodies. n.s = no significance, * P < 0.05 and *** P < 0.001 (Student’s t-test). www.aging-us.com 10652 AGING can bind to AR and suppress the activation of Akt Consistently, xenograft study that investigated the effect pathway in CRPC by downregulating phosphorylation of MYSM1 deficiency in 22Rv1 cells revealed a levels of Akt, c-Raf and GSK-3β. significant increase of tumor volume, suggesting an antitumorigenic role for MYSM1 in CRPC. Moreover, DISCUSSION data mining revealed a significant correlation between MYSM1 and CDK4, CCND3, CCNE1 and RB1, thus In the present study, we identified the dynamic change of providing evidence to support the growth-suppressive MYSM1 expression during PCa transformation and role of MYSM1. progression. We found that MYSM1 levels are downregulated in CRPC tissues compared to localized Despite the repressive function of MYSM1 in cancer primary tumor tissues, indicating a potential role for progression, our findings unveiled distinct molecular MYSM1 in the switch to castration resistance. A series of mechanisms underlying MYSM1-mediated cell growth functional assays demonstrated that MYSM1 is involved inhibition in CRPC. PTEN loss has been implicated as in proliferation and senescence of CRPC cells. Further one cause for castration resistance in both mice and mechanistic investigation indicated that MYSM1-AR human beings [47–49]. Presurgical treatment studies interaction and Akt/c-Raf/GSK-3β pathway may underlie have shown that PTEN inactivation is involved in the role of MYSM1 in CRPC progression. hormone refractoriness after bicalutamide monotherapy [50]. The molecular mechanisms by which PTEN MYSM1 catalyzes deubiquitination of monoubiquitinated deficiency can result in castration resistance are poorly histone H2A at lysine 119 (H2AK119ub) [26]. A understood. A reciprocal negative feedback between growing bank of evidence indicates that dysregulated PI3K/Akt and AR pathways has been implicated to be H2AK119ub levels, which are modulated by various responsible for castration-resistant phenotype displayed deubiquitinases and ubiquitin ligases, are involved in in PTEN-deficient prostate cancers [21, 22]. MYSM1 gene transcription [26, 38], DNA damage repair [39] and has been previously demonstrated to participate in AR replication fork regulation [40], all of which could result transcriptional activity in PTEN null LNCaP cells [26]. in genomic instability and tumorigenesis. Furthermore, Importantly, we found MYSM1 knockdown has no H2A deubiquitinases, such as BAP1, USP14 and USP28, effect on AR expression, suggesting that MYSM1 does have been shown to suppress proliferation and increase not mediate AR target gene transcription via stabilizing radiosensitization in human cancers [41–43]. In contrast, AR protein. The transcriptional action of AR can be previous studies have revealed that ubiquitin ligases, such modulated by cofactors through histone modification, as RNF2 and BMI1, play a pivotal role in tumorigenesis chromatin remodeling or manipulating the interplay and metastasis [44–46]. As a member of deubiquitinase between AR and transcription complex [51, 52]. family, MYSM1 has been reported to be associated with Interestingly, MYSM1-associated gene LCOR has been tumor progression in melanoma and colorectal cancer by established to act as a corepressor of nuclear hormone only few studies with small sample sizes [33, 34]. receptors and inhibit tumor growth and hepatic However, none of the previous reports investigated the lipogenesis via physically interacting with liganded function of MYSM1 in CRPC. Herein, we provided receptors [36, 53]. However, whether MYSM1 also several lines of evidence suggesting a distinct role of function as a AR coregulator is not known. In this work MYSM1 in suppressing CRPC progression. we observed MYSM1-AR interactions which may be Bioinformatics analyses of PCa cohorts from Oncomine also responsible for the role of MYSM1 in androgen- database revealed that MYSM1 is downregulated in responsive gene transcription. Although the CRPC. In addition, the expression of MYSM1 inversely transcriptional activity and localization of AR can be correlates with Gleason grade, Gleason score and regulated by its ubiquitination status, MYSM1 functions recurrence status, establishing MYSM1 as a cancer- as a histone deubiquitinase that is specific for related marker in prostate cancer. Zhu et al. observed a monoubiquitinated H2A. Forming a protein complex decrease of monoubiquitinated H2A (uH2A) levels in 35 with AR may contribute to the recruitment of MYSM1 PCa patient tissues compared to corresponding normal to AR responsive elements. The localization of MYSM1 tissues [26]. Although MYSM1 is functionally essential facilitates alterations of histone modification by for deubiquitination of uH2A, the level of uH2A is coordinating H2A deubiquitination and histone regulated by multifarious deubiquitinases and ubiquitin acetylation, thus leading to transcriptional activation of ligases [26, 43, 45], implying that the decrease of uH2A AR target genes [26]. in PCa may result from a complex process of ubiquitination and deubiquitination. In addition to clinical Given that MYSM1-mediated AR action is ligand- correlation between MYSM1 and cancer progression, our dependent, downregulation of MYSM1, in combination data show that ablating MYSM1 in CRPC cells robustly with androgen deprivation, dramatically impairs AR promotes cell growth and suppresses cellular senescence. signaling in CRPC. Additionally, our findings www.aging-us.com 10653 AGING demonstrated that MYSM1 expression is positively purchased from GenePharma (Shanghai, China). correlated with the levels of PTEN and PHLLPP1 both of Transfection of siRNA duplexes (50 nM) was carried out which are well-established negative regulators of Akt using Lipofectamine 2000 reagent (Invitrogen) signaling. Further investigation in PTEN-deficient C4-2 according to the manufacturer's instructions. Lentiviral cells validated that inhibition of MYSM1 leads to a particles expressing short hairpin RNA (shRNA) significant decrease in PHLPP1 expression. These targeting MYSM1 (shMYSM1) or negative control observations collectively support a link between MYSM1 (shNC) were ordered from Obio Technology (Shanghai, and PI3K/Akt pathway. Repressed AR activity resulting China). Cells were transduced with lentiviral particles from MYSM1 decrease and castration therapy alleviates and selected with medium containing 2 μg/mL inhibitory feedback to Akt signaling. In agreement with puromycin (Sigma-Aldrich) for 1 week to establish the bidirectional crosstalk between PI3K/Akt and AR stable transfections. The detailed sequences for siRNAs pathways, our data show that knockdown of MYSM1 and shRNAs are provided in Supplementary Table 2. results in increased Akt activation, which may potentially account for the malignant growth and development of RNA isolation and quantitative PCR (qPCR) antiandrogen resistance in a setting of castration. Total RNA was extracted from cells used in this study In summary, our results unveil a pivotal role of MYSM1 with TRIzol reagent (Invitrogen) following the in CRPC. Decreased MYSM1 may contribute to manufacturer’s instructions. The RNA samples were androgen-independent growth and castration resistance subjected to reverse transcription reactions using through modulating the reciprocal negative feedback PrimeScript™ RT Master Mix (TakaRa). Quantitative between PI3K/Akt and AR pathways. These findings real-time polymerase chain reaction (qRT-PCR) suggest a potential role of therapeutic targets and analyses were then performed using SYBR® Premix biomarkers for MYSM1 in castration-resistant prostate ExTaq™ II (TaKaRa) to determine the expression levels cancer. of resulting cDNA based on the manufacturer’s protocols. Beta-actin was used for an internal control. MATERIALS AND METHODS The relative abundance of mRNA was calculated via the comparative Ct method after normalization. The primer Prostate cancer samples pairs used for qRT-PCR analyses are listed in Supplementary Table 3. Paraffin-embedded human prostate cancer specimens were obtained from patients (n = 15) who had undergone Western blotting radical prostatectomy and benign prostatic hyperplasia (BPH) specimens (n = 13) were collected by transurethral Proteins for Immunoblotting were extracted using RIPA resection of prostate in Tangdu Hospital affiliated to lysis buffer containing protease and phosphatase Fourth Military Medical University (FMMU), with inhibitors. The concentrations of extracted proteins were written informed consents from all patients. Collected quantified using BCA assay (Thermo). Reduced proteins samples were pathologically evaluated. Details of PCa were separated by SDS/PAGE gel and then transferred patients were described in Supplementary Table 1. All to polyvinylidene fluoride (PVDF) membrane experiments were approved by the Medical Ethics (Millipore). After blocked in 5% bovine serum albumin Committee of Tangdu Hospital (TDLL-201504-12). (BSA) for 1h at room temperature, the PVDF membrane was incubated with indicated primary antibody overnight Cell culture at 4 °C. HRP-conjugated secondary antibodies (Jackson ImmunoResearch) and ECL substrate (Millipore) were The CRPC cell lines C4-2 and 22Rv1 were obtained used for signal detection on FluorChem FC2 system from American Type Culture Collection (ATCC). All (Alpha Innotech, San Leandro, USA) according to the cell lines were grown in Roswell Park Memorial manufacturer’s instructions. Detailed information of the Institute medium-1640 (RPMI-1640, Gibco) with 10% antibodies is provided in Supplementary Table 4. charcoal/dextran-treated FBS (Bioind) and 1% penicillin-streptomycin (Gibco). All cells were incubated Cell proliferation assay at 37 °C in humidified 5% CO2 atmosphere. Cell proliferation was measured using MTT assay kit Oligonucleotides transfection and lentiviral (Sigma-Aldrich) according to instructions of the transduction manufacturer. Briefly, Cells were seeded in triplicate at 1 × 103/well in 96-well plates. Assay was performed for All synthetic small interfering RNA (siRNA) against indicated time point by adding 20μL MTT solution to MYSM1 (siMYSM1) or negative control (NC) were the medium. Then the plate was incubated at 37 °C. www.aging-us.com 10654 AGING After incubation for 4h, dimethyl sulfoxide (DMSO) was citrate buffer (pH 6.0) for 30 min. After inactivation of used to dissolve the formazan crystals, followed by endogenous peroxidases (10 min, 3% H2O2) and reading absorbance at 490 nm with a microplate reader blocking of nonspecific binding (1 h, Beyotime Immunol (Bio-Rad). Each experiment was repeated at least three Staining Blocking Buffer), the sections were then times. incubated with diluted primary antibodies overnight at 4 °C, followed by sequential incubation with biotinylated Cell cycle and apoptosis analyses secondary antibody (10 min, Maxim, Fuzhou, China) and streptavidin horseradish peroxidase (10 min, HRP, For cell cycle analysis, cells were harvested and fixed in Maxim) at room temperature. Antigen binding was 70% ethanol overnight at 4 °C. Then cells were treated visualized using standard DAB (Maxim) staining and with staining solution containing RNase A (100 μg/mL) haematoxylin counterstaining. Images were taken under and propidium iodide (PI, 50 μg/mL) at 37 °C for 1 h a light microscope. Detailed information of primary and subjected to DNA content analysis using FACS scan antibodies is provided in Supplementary Table 4. flow cytometer (BD Biosciences). For apoptosis analysis, cells were treated with siRNA transfection and Co-immunoprecipitation (Co-IP) assay collected for apoptosis assay at 48h after treatment. Apoptotic cells were labeled with Annexin V-FITC and Cells were lysed with IP buffer (1% NP-40, 20 mM Tris- PI (BD Biosciences) and quantified by CYTOMICS FC HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 0.25% 500 flow cytometer (Beckman Coulter). sodium deoxycholate and protease inhibitor cocktail). Cellular extracts (1 mg) were pre-cleared with Protein Senescence-associated β-galactosidase (SA-β-gal) A/G PLUS-Agarose beads (sc-2003, Santa Cruz staining assay Biotechnology), followed by overnight incubation with 3 μg of anti-IgG, MYSM1 or AR antibody on a rotator at 4 Cellular senescence was assessed through detecting the °C. Then Protein A/G PLUS-Agarose beads were added activity of β-galactosidase using a SA-β-gal staining kit for 2 h incubation to precipitate immune complexes. The (Beyotime) according to the manufacturer’s instructions. beads were washed five times with IP buffer and boiled Briefly, cultured cells in 6-well plates were immersed in for 5 min in protein loading buffer. Finally, the fixative solution for 15 minutes at room temperature. immunoprecipitates were subjected to Western blot After rinse with PBS, cells were incubated with freshly analysis. Detailed information of the antibodies is prepared staining work solution overnight at 37 °C. provided in Supplementary Table 4. Then, stained cells were photographed and counted using a light microscope. Bioinformatics analysis

In vivo tumorigenicity assay The expression levels and clinical significance of MYSM1 in prostate cancer and other tumors were The animal studies were approved by Institutional analyzed by Gene Expression Profiling Interactive Animal Care and Use Committee of FMMU and Analysis (GEPIA, http://gepia.cancer-pku.cn/index.html) conducted in accordance with ethical regulations and [54] and Oncomine (http://www.oncomine.org/). humane treatment. To evaluate the tumorigenicity in Identification of MYSM1-associated genes and gene vivo, male BALB/c nude mice (6-week-old, obtained expression correlation analysis was performed via the from the Experimental Animal Center of FMMU, n = LinkedOmics database (http://www.linkedomics.org/) 3/group) were inoculated subcutaneously with 5 × 106 [55]. Genomic associations and network analyses of cells (22Rv1/shMYSM1 or 22Rv1/shNC) suspended in MYSM1 were carried out using Cancer Regulome 100μL PBS. Tumor size was monitored every 6 days and Datasets (http://explorer.cancerregulome.org/). The the volume was estimated using calipers (volume = Web-based Gene SeT AnaLysis Toolkit (WebGestalt, length × width2/2). Thirty days after inoculation, mice http://www.webgestalt.org/) was used to perform GO were sacrificed. The tumors were weighted and and KEGG pathway analyses. subjected to histological analysis. Statistical analysis Immunohistochemistry (IHC) Statistical analysis was carried out using SPSS 18.0 The formalin-fixed paraffin-embedded (FFPE) sections statistical software (SPSS, IBM Corporation). Data (5 μm) were stained with hematoxylin–eosin for tumor derived from three separate experiments are shown as morphology visualization. For IHC staining, xylene- mean ± standard deviation (SD) or mean ± standard deparaffinized and rehydrated sections were subjected to error of mean (SEM). Statistical analysis of two heat-mediated antigen retrieval with microwave in independent groups was determined by Student’s t-test

www.aging-us.com 10655 AGING (two-sided), except for gene expression correlation receptors in prostate cancer: a hard habit to break? analysis that employed a Pearson correlation coefficient. Cancer Cell. 2009; 16:458–62. For comparison of more than two groups, one-way https://doi.org/10.1016/j.ccr.2009.11.006 ANOVA test was used. Survival analyses were PMID:19962664 performed by Kaplan-Meier survival curve and 6. Montgomery RB, Mostaghel EA, Vessella R, Hess DL, determined using the Log-rank test. Statistically Kalhorn TF, Higano CS, True LD, Nelson PS. significant differences (* P < 0.05, ** P < 0.01, *** P < Maintenance of intratumoral androgens in metastatic 0.001 and **** P < 0.0001) were indicated. prostate cancer: a mechanism for castration-resistant

tumor growth. Cancer Res. 2008; 68:4447–54. ACKNOWLEDGMENTS https://doi.org/10.1158/0008-5472.CAN-08-0249 PMID:18519708 We thank the help from State Key Laboratory of Cancer Biology, Department of Immunology, Fourth Military 7. Roudier MP, True LD, Higano CS, Vesselle H, Ellis W, Medical University. Lange P, Vessella RL. Phenotypic heterogeneity of end- stage prostate carcinoma metastatic to bone. Hum CONFLICTS OF INTEREST Pathol. 2003; 34:646–53. https://doi.org/10.1016/S0046-8177(03)00190-4 The authors declare no conflicts of interest. PMID:12874759 8. Fruman DA, Rommel C. PI3K and cancer: lessons, FUNDING challenges and opportunities. Nat Rev Drug Discov. 2014; 13:140–56. The study was supported by the National Natural Science https://doi.org/10.1038/nrd4204 Foundation of China (81772745) and the Science and PMID:24481312 Technology Innovation Development Foundation of Fourth Military Medical University (2016XB078). 9. Liu P, Cheng H, Roberts TM, Zhao JJ. Targeting the phosphoinositide 3-kinase pathway in cancer. Nat Rev REFERENCES Drug Discov. 2009; 8:627–44. https://doi.org/10.1038/nrd2926

PMID:19644473 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018; 68:7–30. 10. Gao S, Zhao Z, Wu R, Wu L, Tian X, Zhang Z. MiR-146b https://doi.org/10.3322/caac.21442 inhibits autophagy in prostate cancer by targeting the PMID:29313949 PTEN/Akt/mTOR signaling pathway. Aging (Albany NY). 2018; 10:2113–21. 2. Tran C, Ouk S, Clegg NJ, Chen Y, Watson PA, Arora V, https://doi.org/10.18632/aging.101534 Wongvipat J, Smith-Jones PM, Yoo D, Kwon A, PMID:30153654 Wasielewska T, Welsbie D, Chen CD, et al. Development of a second-generation antiandrogen for 11. Kaarbø M, Mikkelsen OL, Malerød L, Qu S, Lobert VH, treatment of advanced prostate cancer. Science. 2009; Akgul G, Halvorsen T, Maelandsmo GM, Saatcioglu F. 324:787–90. PI3K-AKT-mTOR pathway is dominant over androgen https://doi.org/10.1126/science.1168175 receptor signaling in prostate cancer cells. Cell Oncol. PMID:19359544 2010; 32:11–27. https://doi.org/10.3233/CLO-2009-0487 3. de Bono JS, Logothetis CJ, Molina A, Fizazi K, North S, PMID:20203370 Chu L, Chi KN, Jones RJ, Goodman OB Jr, Saad F, Staffurth JN, Mainwaring P, Harland S, et al, and COU- 12. Wang N, Yao M, Xu J, Quan Y, Zhang K, Yang R, Gao AA-301 Investigators. Abiraterone and increased WQ. Autocrine Activation of CHRM3 Promotes survival in metastatic prostate cancer. N Engl J Med. Prostate Cancer Growth and Castration Resistance via 2011; 364:1995–2005. CaM/CaMKK-Mediated Phosphorylation of Akt. Clin https://doi.org/10.1056/NEJMoa1014618 Cancer Res. 2015; 21:4676–85. PMID:21612468 https://doi.org/10.1158/1078-0432.CCR-14-3163 PMID:26071486 4. Cai Z, Wu Y, Li Y, Ren J, Wang L. BCAR4 activates GLI2 signaling in prostate cancer to contribute to castration 13. Edlind MP, Hsieh AC. PI3K-AKT-mTOR signaling in resistance. Aging (Albany NY). 2018; 10:3702–12. prostate cancer progression and androgen deprivation https://doi.org/10.18632/aging.101664 therapy resistance. Asian J Androl. 2014; 16:378–86. PMID:30513511 https://doi.org/10.4103/1008-682X.122876 PMID:24759575 5. Attard G, Cooper CS, de Bono JS. Steroid hormone www.aging-us.com 10656 AGING 14. Taylor BS, Schultz N, Hieronymus H, Gopalan A, Xiao Y, 22. Carver BS, Chapinski C, Wongvipat J, Hieronymus H, Carver BS, Arora VK, Kaushik P, Cerami E, Reva B, Chen Y, Chandarlapaty S, Arora VK, Le C, Koutcher J, Antipin Y, Mitsiades N, Landers T, et al. Integrative Scher H, Scardino PT, Rosen N, Sawyers CL. Reciprocal genomic profiling of human prostate cancer. Cancer feedback regulation of PI3K and androgen receptor Cell. 2010; 18:11–22. signaling in PTEN-deficient prostate cancer. Cancer https://doi.org/10.1016/j.ccr.2010.05.026 Cell. 2011; 19:575–86. PMID:20579941 https://doi.org/10.1016/j.ccr.2011.04.008 PMID:21575859 15. McMenamin ME, Soung P, Perera S, Kaplan I, Loda M, Sellers WR. Loss of PTEN expression in paraffin- 23. Toren P, Kim S, Cordonnier T, Crafter C, Davies BR, Fazli embedded primary prostate cancer correlates with L, Gleave ME, Zoubeidi A. Combination AZD5363 with high Gleason score and advanced stage. Cancer Res. Enzalutamide Significantly Delays Enzalutamide- 1999; 59:4291–96. PMID:10485474 resistant Prostate Cancer in Preclinical Models. Eur Urol. 2015; 67:986–90. 16. Lotan TL, Gurel B, Sutcliffe S, Esopi D, Liu W, Xu J, Hicks JL, Park BH, Humphreys E, Partin AW, Han M, https://doi.org/10.1016/j.eururo.2014.08.006 PMID:25151012 Netto GJ, Isaacs WB, De Marzo AM. PTEN protein loss by immunostaining: analytic validation and 24. Marques RB, Aghai A, de Ridder CM, Stuurman D, prognostic indicator for a high risk surgical cohort of Hoeben S, Boer A, Ellston RP, Barry ST, Davies BR, prostate cancer patients. Clin Cancer Res. 2011; Trapman J, van Weerden WM. High Efficacy of 17:6563–73. Combination Therapy Using PI3K/AKT Inhibitors with https://doi.org/10.1158/1078-0432.CCR-11-1244 Androgen Deprivation in Prostate Cancer Preclinical PMID:21878536 Models. Eur Urol. 2015; 67:1177–85. https://doi.org/10.1016/j.eururo.2014.08.053 17. Cancer Genome Atlas Research Network. The PMID:25220373 Molecular Taxonomy of Primary Prostate Cancer. Cell. 2015; 163:1011–25. 25. Kato M, Banuelos CA, Imamura Y, Leung JK, Caley DP, https://doi.org/10.1016/j.cell.2015.10.025 Wang J, Mawji NR, Sadar MD. Cotargeting Androgen PMID:26544944 Receptor Splice Variants and mTOR Signaling Pathway 18. Jiao J, Wang S, Qiao R, Vivanco I, Watson PA, Sawyers for the Treatment of Castration-Resistant Prostate Cancer. Clin Cancer Res. 2016; 22:2744–54. CL, Wu H. Murine cell lines derived from Pten null https://doi.org/10.1158/1078-0432.CCR-15-2119 prostate cancer show the critical role of PTEN in PMID:26712685 hormone refractory prostate cancer development. Cancer Res. 2007; 67:6083–91. 26. Zhu P, Zhou W, Wang J, Puc J, Ohgi KA, Erdjument- https://doi.org/10.1158/0008-5472.CAN-06-4202 Bromage H, Tempst P, Glass CK, Rosenfeld MG. A PMID:17616663 histone H2A deubiquitinase complex coordinating histone acetylation and H1 dissociation in 19. Wang S, Gao J, Lei Q, Rozengurt N, Pritchard C, Jiao J, transcriptional regulation. Mol Cell. 2007; 27:609–21. Thomas GV, Li G, Roy-Burman P, Nelson PS, Liu X, Wu H. Prostate-specific deletion of the murine Pten tumor https://doi.org/10.1016/j.molcel.2007.07.024 PMID:17707232 suppressor gene leads to metastatic prostate cancer. Cancer Cell. 2003; 4:209–21. 27. Belle JI, Petrov JC, Langlais D, Robert F, Cencic R, Shen https://doi.org/10.1016/S1535-6108(03)00215-0 S, Pelletier J, Gros P, Nijnik A. Repression of p53-target PMID:14522255 gene Bbc3/PUMA by MYSM1 is essential for the survival of hematopoietic multipotent progenitors and 20. Wang Y, Kreisberg JI, Ghosh PM. Cross-talk between the androgen receptor and the phosphatidylinositol 3- contributes to stem cell maintenance. Cell Death kinase/Akt pathway in prostate cancer. Curr Cancer Differ. 2016; 23:759–75. https://doi.org/10.1038/cdd.2015.140 PMID:26768662 Drug Targets. 2007; 7:591–604. https://doi.org/10.2174/156800907781662248 28. Belle JI, Langlais D, Petrov JC, Pardo M, Jones RG, Gros PMID:17896924 P, Nijnik A. p53 mediates loss of hematopoietic stem cell function and lymphopenia in Mysm1 deficiency. 21. Mulholland DJ, Tran LM, Li Y, Cai H, Morim A, Wang S, Plaisier S, Garraway IP, Huang J, Graeber TG, Wu H. Cell Blood. 2015; 125:2344–48. autonomous role of PTEN in regulating castration- https://doi.org/10.1182/blood-2014-05-574111 PMID:25710881 resistant prostate cancer growth. Cancer Cell. 2011; 19:792–804. 29. Nijnik A, Clare S, Hale C, Raisen C, McIntyre RE, Yusa K, https://doi.org/10.1016/j.ccr.2011.05.006 Everitt AR, Mottram L, Podrini C, Lucas M, Estabel J, PMID:21620777 Goulding D, Adams N, et al, and Sanger Institute www.aging-us.com 10657 AGING Microarray Facility, and Sanger Mouse Genetics 37. Gubanova E, Issaeva N, Gokturk C, Djureinovic T, Project. The critical role of histone H2A-deubiquitinase Helleday T. SMG-1 suppresses CDK2 and tumor growth Mysm1 in hematopoiesis and lymphocyte by regulating both the p53 and Cdc25A signaling differentiation. Blood. 2012; 119:1370–79. pathways. Cell Cycle. 2013; 12:3770–80. https://doi.org/10.1182/blood-2011-05-352666 https://doi.org/10.4161/cc.26660 PMID:22184403 PMID:24107632 30. Gatzka M, Tasdogan A, Hainzl A, Allies G, Maity P, 38. Yu H, Mashtalir N, Daou S, Hammond-Martel I, Ross J, Wilms C, Wlaschek M, Scharffetter-Kochanek K. Sui G, Hart GW, Rauscher FJ 3rd, Drobetsky E, Milot E, Interplay of H2A deubiquitinase 2A-DUB/Mysm1 and Shi Y, Affar B. The ubiquitin carboxyl hydrolase BAP1 the p19(ARF)/p53 axis in hematopoiesis, early T-cell forms a ternary complex with YY1 and HCF-1 and is a development and tissue differentiation. Cell Death critical regulator of gene expression. Mol Cell Biol. Differ. 2015; 22:1451–62. 2010; 30:5071–85. https://doi.org/10.1038/cdd.2014.231 https://doi.org/10.1128/MCB.00396-10 PMID:25613381 PMID:20805357 31. Panda S, Nilsson JA, Gekara NO. Deubiquitinase 39. Yu H, Pak H, Hammond-Martel I, Ghram M, Rodrigue A, MYSM1 Regulates Innate Immunity through Daou S, Barbour H, Corbeil L, Hébert J, Drobetsky E, Inactivation of TRAF3 and TRAF6 Complexes. Masson JY, Di Noia JM, Affar B. Tumor suppressor and Immunity. 2015; 43:647–59. deubiquitinase BAP1 promotes DNA double-strand https://doi.org/10.1016/j.immuni.2015.09.010 break repair. Proc Natl Acad Sci USA. 2014; 111:285– PMID:26474655 90. 32. Panda S, Gekara NO. The deubiquitinase MYSM1 https://doi.org/10.1073/pnas.1309085110 PMID:24347639 dampens NOD2-mediated inflammation and tissue damage by inactivating the RIP2 complex. Nat 40. Lee HS, Lee SA, Hur SK, Seo JW, Kwon J. Stabilization Commun. 2018; 9:4654. and targeting of INO80 to replication forks by BAP1 https://doi.org/10.1038/s41467-018-07016-0 during normal DNA synthesis. Nat Commun. 2014; PMID:30405132 5:5128. https://doi.org/10.1038/ncomms6128 PMID:25283999 33. Wilms C, Kroeger CM, Hainzl AV, Banik I, Bruno C, Krikki I, Farsam V, Wlaschek M, Gatzka MV. 41. Zhang Y, Zhuang L, Gan B. BAP1 suppresses tumor MYSM1/2A-DUB is an epigenetic regulator in human development by inducing ferroptosis upon SLC7A11 melanoma and contributes to tumor cell growth. repression. Mol Cell Oncol. 2018; 6:1536845. Oncotarget. 2017; 8:67287–99. https://doi.org/10.1080/23723556.2018.1536845 https://doi.org/10.18632/oncotarget.18617 PMID:30788415 PMID:28978033 42. Sharma A, Alswillah T, Singh K, Chatterjee P, Willard B, 34. Li Y, Li J, Liu H, Liu Y, Cui B. Expression of MYSM1 is Venere M, Summers MK, Almasan A. USP14 regulates associated with tumor progression in colorectal cancer. DNA damage repair by targeting RNF168-dependent PLoS One. 2017; 12:e0177235. ubiquitination. Autophagy. 2018; 14:1976–90. https://doi.org/10.1371/journal.pone.0177235 https://doi.org/10.1080/15548627.2018.1496877 PMID:28498834 PMID:29995557 35. Alessio N, Squillaro T, Özcan S, Di Bernardo G, Venditti 43. Li F, Han H, Sun Q, Liu K, Lin N, Xu C, Zhao Z, Zhao W. M, Melone M, Peluso G, Galderisi U. Stress and stem USP28 regulates deubiquitination of histone H2A and cells: adult Muse cells tolerate extensive genotoxic cell proliferation. Exp Cell Res. 2019; 379:11–18. stimuli better than mesenchymal stromal cells. https://doi.org/10.1016/j.yexcr.2019.03.026 Oncotarget. 2018; 9:19328–41. PMID:30910399 https://doi.org/10.18632/oncotarget.25039 44. Zhu S, Zhao D, Yan L, Jiang W, Kim JS, Gu B, Liu Q, PMID:29721206 Wang R, Xia B, Zhao JC, Song G, Mi W, Wang RF, et al. 36. Asim M, Hafeez BB, Siddiqui IA, Gerlach C, Patz M, BMI1 regulates androgen receptor in prostate cancer Mukhtar H, Baniahmad A. Ligand-dependent independently of the polycomb repressive complex 1. corepressor acts as a novel androgen receptor Nat Commun. 2018; 9:500. corepressor, inhibits prostate cancer growth, and is https://doi.org/10.1038/s41467-018-02863-3 functionally inactivated by the Src protein kinase. J Biol PMID:29402932 Chem. 2011; 286:37108–17. 45. Rai K, Akdemir KC, Kwong LN, Fiziev P, Wu CJ, Keung https://doi.org/10.1074/jbc.M111.292771 EZ, Sharma S, Samant NS, Williams M, Axelrad JB, PMID:21856747 www.aging-us.com 10658 AGING Shah A, Yang D, Grimm EA, et al. Dual Roles of RNF2 50. Ham WS, Cho NH, Kim WT, Ju HJ, Lee JS, Choi YD. in Melanoma Progression. Cancer Discov. 2015; Pathological effects of prostate cancer correlate with 5:1314–27. neuroendocrine differentiation and PTEN expression https://doi.org/10.1158/2159-8290.CD-15-0493 after bicalutamide monotherapy. J Urol. 2009; PMID:26450788 182:1378–84. https://doi.org/10.1016/j.juro.2009.06.025 46. Su WJ, Fang JS, Cheng F, Liu C, Zhou F, Zhang J. PMID:19683286 RNF2/Ring1b negatively regulates p53 expression in selective cancer cell types to promote tumor 51. Heemers HV, Tindall DJ. Androgen receptor (AR) development. Proc Natl Acad Sci USA. 2013; coregulators: a diversity of functions converging on 110:1720–25. and regulating the AR transcriptional complex. Endocr https://doi.org/10.1073/pnas.1211604110 Rev. 2007; 28:778–808. PMID:23319651 https://doi.org/10.1210/er.2007-0019 PMID:17940184 47. Gao H, Ouyang X, Banach-Petrosky WA, Shen MM, 52. Chang CY, McDonnell DP. Androgen receptor-cofactor Abate-Shen C. Emergence of androgen independence interactions as targets for new drug discovery. Trends at early stages of prostate cancer progression in Pharmacol Sci. 2005; 26:225–28. Nkx3.1; Pten mice. Cancer Res. 2006; 66:7929–33. https://doi.org/10.1016/j.tips.2005.03.002 https://doi.org/10.1158/0008-5472.CAN-06-1637 PMID:15860367 PMID:16912166 53. Song Y, Shan S, Zhang Y, Liu W, Ding W, Ren W, Xia H, 48. Zhang W, Zhu J, Efferson CL, Ware C, Tammam J, Li X, Zhang Q, Zhao L, Li X, Yan J, Ying H. Ligand- Angagaw M, Laskey J, Bettano KA, Kasibhatla S, Reilly dependent corepressor acts as a novel corepressor of JF, Sur C, Majumder PK. Inhibition of tumor growth thyroid hormone receptor and represses hepatic progression by antiandrogens and mTOR inhibitor in a lipogenesis in mice. J Hepatol. 2012; 56:248–54. Pten-deficient mouse model of prostate cancer. Cancer https://doi.org/10.1016/j.jhep.2011.07.014 Res. 2009; 69:7466–72. PMID:21827732 https://doi.org/10.1158/0008-5472.CAN-08-4385 54. Tang Z, Li C, Kang B, Gao G, Li C, Zhang Z. GEPIA: a web PMID:19738074 server for cancer and normal gene expression profiling 49. Ferraldeschi R, Nava Rodrigues D, Riisnaes R, Miranda and interactive analyses. Nucleic Acids Res. 2017; S, Figueiredo I, Rescigno P, Ravi P, Pezaro C, Omlin A, 45:W98–102. Lorente D, Zafeiriou Z, Mateo J, Altavilla A, et al. PTEN https://doi.org/10.1093/nar/gkx247 PMID:28407145 protein loss and clinical outcome from castration- 55. Vasaikar SV, Straub P, Wang J, Zhang B. LinkedOmics: resistant prostate cancer treated with abiraterone analyzing multi-omics data within and across 32 cancer acetate. Eur Urol. 2015; 67:795–802. types. Nucleic Acids Res. 2018; 46:D956–63. https://doi.org/10.1016/j.eururo.2014.10.027 https://doi.org/10.1093/nar/gkx1090 PMID:25454616 PMID:29136207

www.aging-us.com 10659 AGING SUPPLEMENTARY MATERIALS

Supplementary Figures

Supplementary Figure 1. MYSM1 is downregulated in numerous human tumors, including ACC, CESC, COAD, LUAD (A), LUSC, OV, READ, SKCM (B), THCA, UCEC and UCS (C). Data were acquired from TCGA database and analyzed via GEPIA bioinformatics.

www.aging-us.com 10660 AGING

Supplementary Figure 2. Upregulated MYSM1 in DLBC (A) and THYM (B). Data were acquired from TCGA database and analyzed via GEPIA bioinformatics.

Supplementary Figure 3. Flow cytometry analysis of cell apoptosis in C4-2/22Rv1 cells treated with NC/siMYSM1. Data are shown as mean ± SD of 3 replicates. n.s = no significance. www.aging-us.com 10661 AGING Supplementary Tables

Supplementary Table 1. Information on PCa patient tissues collected for the study. Patient number ID Number Gender Age PSA (ng/ml) Gleason score 1 14876718 Male 72 24.74 5+3 2 14718613 Male 67 41.47 4+3 3 14647229 Male 71 24.99 4+3 4 14644672 Male 65 31.97 4+3 5 14483052 Male 65 13.96 3+3 6 14398145 Male 82 0.754 3+4 7 13624506 Male 66 25.18 4+3 8 11919601 Male 71 14.7 3+4 9 14931329 Male 53 43.25 4+3 10 14107451 Male 74 38.7 4+3 11 14076409 Male 70 13.41 3+4 12 14845766 Male 67 9.77 4+3 13 14233999 Male 72 19.46 5+4

Supplementary Table 2. Sequences of the oligonucleotides in this study. Oligonucleotides Sense (5′-3′)/Target sequence Antisense (5′-3′) siNC UUCUCCGAACGUGUCACGUTT ACGUGACACGUUCGGAGAATT siMYSM1 CCGGCCAUAAUCUUCAAGUTT ACUUGAAGAUUAUGGCCGGTT shNC TTCTCCGAACGTGTCACGT shMYSM1#1 CCAATCAAGGAGAATTCAT shMYSM1#2 CCAGAACAGGAAATAGAAA

Supplementary Table 3. Primers used for qRT-PCR analyses in this study. Gene Forward (5′-3′) Reverse (5′-3′) MYSM1 CACAGGTACCCACATTGCTG CTGTATCATAGGCCCCCTCA AR ATTGCGAGAGAGCTGCATCA GGGCACTTGCACAGAGATGA PHLPP1 GCAGGAAAACCTCACAGCA AGGATGACTTGGCGTCTTGT β-actin TGGCATCCACGAAACTACC GTGTTGGCGTACAGGTCTT

www.aging-us.com 10662 AGING Supplementary Table 4. Antibodies used for Western blot analysis, Co-immunoprecipitation and Immunohistochemistry in this study. Antigen Species Applications and dilutions Source IgG Rabbit 3µg in Co-IP Cell Signaling Technology #2729 MYSM1 Rabbit WB (1:1000); Abcam # ab193081 3µg in Co-IP MYSM1 Rabbit IHC (1:60) Sangon Biotech #D261010 AR Rabbit WB (1:1000); IHC (1:250); Abcam # ab108341 3µg in Co-IP Akt(pan) Rabbit WB (1:1000) Cell Signaling Technology #4691 p-Akt(Ser473) Rabbit WB (1:1000); IHC (1:100) Cell Signaling Technology #4060 p-Akt(Thr308) Rabbit WB (1:1000) Cell Signaling Technology #13038 c-Raf Rabbit WB (1:1000) Cell Signaling Technology #9422 p-c-Raf(Ser259) Rabbit WB (1:1000) Cell Signaling Technology #9421 GSK-3β Rabbit WB (1:1000) Cell Signaling Technology #9315 p-GSK-3β(Ser9) Rabbit WB (1:1000) Cell Signaling Technology #5558 p-PDK1(Ser241) Rabbit WB (1:1000) Cell Signaling Technology #3438 Ki-67 Mouse IHC (1:400) Cell Signaling Technology #9449 β-actin Mouse WB (1:2000) Sigma-Aldrich A1978 Anti-rabbit IgG Goat WB (1:5000) Jackson ImmunoResearch #124791 Anti-mouse IgG Goat WB (1:5000) Genshare Biological #JC-PB002H